UC Berkeley

Many psychophysics investigations have implemented pre-cues to direct an observer's attention to a specific location or feature. There is controversy over the mechanisms of involuntary attention and whether perceptual or decision processes can enhance target detection and identification as measured by accuracy judgments. Through four main experiments, this dissertation research has indicated that both involuntary and voluntary attention improve target identification and localization accuracy even when cues are non-predictive. The first experiment was conducted to assess the validity of the mask-dependent cueing hypothesis and to determine if involuntary attention improves target identification accuracy. A two-alternative force choice experimental procedure using the method of constant stimuli was conducted using non-predictive cues and both masked and unmasked target stimuli. The results indicated that involuntary attention improved target identification accuracy for unmasked stimuli across the entire Weibull psychometric function. The second experiment introduced multinomial modeling of observed data to assess the extent of response bias which has been shown to confound cueing experiments. In a seven-alternative force choice experiment, observers reported both the location and identification of masked stimuli presented across a range of temporal parameters spanning the time course of both voluntary and involuntary attention. The multinomial modeling removed the response bias and the results indicated a strong cueing effect for both voluntary and involuntary spatial attention.

The third experiment used the same multinomial modeling technique to remove response bias, but stimuli were unmasked and 6 stimulus contrast levels were tested ranging from 19% to 100% contrast. Results indicated strong cueing effects across the entire psychometric function. The fourth experiment was a six-alternative force choice feature-attention task in which observers reported the identity and location of target stimuli following a feature cue. The results indicated that under both involuntary and voluntary attention, response accuracy was increased. The combined results indicate that accuracy is enhanced with voluntary and involuntary attention for both feature-based and visuo-spatial attention using non-predictive cues.